Identifying Reticent Respondents: Assessing the Quality of Survey Data on Corruption and Values

Randomized response methods, which were designed to elicit candid answers to sensitive questions, have not succeeded in eliminating reticence in survey responses. We implement a methodology that effectively stands the randomized response technique on its head, using it to identify reticent respondents. In a sample of Romanian company officials, we identify a specific 10% of respondents as reticent with near certainty and estimate that roughly 40% of the whole sample were actually reticent. The identifiably reticent respondents admit to corruption interactions significantly less often than others do. They are also more likely to state that it is impermissible to break socially beneficial rules. We show that reticence is related to the respondent's age and the colonial heritage of the respondent's region. These results suggest some difficulties in making cross-country comparisons of corruption and of values using the types of survey data often employed in social science research and policy analysis.